[Pw_forum] Slow calculation of (5, 5) CNT transmission and complex bands.

[chengyu yang]

Thank you Dr.Sclauzero and all,

     I've changed the pseudo potential to USPP, set the ecutwfc down to 35,
and reduce theepsproj=1.d-5, and I ran the whole procedure to calculate the
band structure of my (5,5) CNT again.

   This time the procedure runs well and faster  until I got to the bands.x
step, I got " negative rho (up, down):  0.502E-03 0.000E+00" at the
bands.out, and it stopped there. I couldn't proceed further.

    My another question is about the K points, I could generate a k point
path via Xcrysden, but I couldn't generate the all the k points I need to
input manually for the 'nscf'  input as in the examples. I defined 50 k
points in Xcrysden, however, the final result was only the few high symmetry
points I defined. I don't know how to get all the k points the 'nscf' needs?
I've tried the kpoints.x, but I don't know how the mesh n1,n2,n3 can be
connected to the k point path by Xcrysden, and how I should set up them.

      Thanks beforehand for anyone who can explain these things to me, I
really appreciate it.

      Best regards.

Chengyu Yang
Material Science, University of Central Florida
USA






2011/8/2 Gabriele Sclauzero <gabriele.sclauzero at epfl.ch>

>
> Il giorno 29/lug/2011, alle ore 21.16, chengyu yang ha scritto:
>
> Dear All,
>          After two hours , my transmission calculation comes out a result.
>
>
> And you couldn't wait for a couple of hours before writing to the forum
> saying that you were in trouble?...
>
> Assuming that your PWscf calculation has been converged, the important
> thing here is to look at the basis set reduction performed by PWcond:
>
>  ngper, shell number =         2809         299
>  ngper, n2d =         2809         472
>
> It looks to me that the n2d could be too small compared to the original
> size of the basis set (ngper). As I wrote in my previous email, you should
> try to decrease epsproj and check that the results (i.e. the complex bands)
> do not change appreciably. The eigenvalues at real k from PWcond need to be
> on the top of the band structure computed with PWscf.
> Also compare with previous theoretical results, I guess that there should
> be several works about the conductance of CNTs.
> cuplot is needed only to plot the scattering states. I guess that you do
> not need them at this stage.
>
> HTH
>
> GS
>
> But I thought it may be not converged and didn't get the final result.
> May anybody have a look at my output file? I attached in this email.
>        My input file is like this:
>        &inputcond
>     prefixl='55'
>     tran_file='trans.cnt55',
>     ikind=0
>     energy0=4.d0
>     denergy=-0.1d0
>     ewind=3.d0
>     epsproj=1.d-4
>     delgep = 1.d-9
>     cutplot = 3.d0
>  /
>     1
>     0.0 0.0 1.0
>     5
>
>
> Is there somebody that can give me an advice?
> I really appreciate it.
>
> Chengyu Yang
> Materials Science and Engineering,
> University of Central Florida, USA
>
> <transmission calculation>_______________________________________________
>
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
>
>
>
> § Gabriele Sclauzero, EPFL SB ITP CSEA
> *   PH H2 462, Station 3, CH-1015 Lausanne*
>
>
>
>
>
>
>
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20110808/2e06099f/attachment.html>